Rigol DS2000A calibration problem

[Willi100]

Hello,

I have a DS2072A with upgrade to DS2302 (300MHz)
After upgrade I recognized that the low frequency and DC performance is not good as before. Therefore I did the Self Calibration. But the calibration did not help. The calibration makes it worse.

For example:
Test Signal: 1 MHz Sinus with 0dBm = 224mV RMS
Measured with Boonton RF- Millivoltmeter 92E: 224mV RMS
Measured with DS2072A before upgrade:         224mV RMS (@500ns)
Measured with DS2072A after upgrade:           227mV RMS (@500ns)
After calibration I measured:                           233mV RMS (@500ns)

My Questions:
1. Is there any way to get the factory calibration back ? (Factory Reset....)
2. Is there a possibility to control the internal (DC) voltage reference?
3. Exists any  Service Mode or Service commands ?
Or any other ideas ?

---------------------------------------------------
Test gear:
Signal Generator IFR2023B
TRUE RMS  RF Level Meter RACAL DANA 9303
BOONTON RF Millivoltmeter 92E
---------------------------------------------------

Many thanks for help
Willi

[David Hess]

What are the results if averaging is used instead of sampling?  What about if bandwidth limiting is used?

I would assume that with the higher bandwidth, more high frequency noise is being included in the oscilloscope's RMS measurement.

The discrepancy between the Boonton and Rigol looks like it is within the combined accuracy of the two instruments.  Just the Boonton by itself should only be within ( ( 300mV + 224mV ) * 1% ) = 5.24mV and the Rigol is worse than that.

The Rigol's manual is depressingly sparse on the subject of measurement accuracy.

[Willi100]

Hi David,

Averaging  or Normal: No difference
With bandwidth limit: No difference (some times 1mV)
I checked the measurement also with my Racal Dana TRUE  Level Meter.
The difference is stable. (I know both meters are accurate)
It is not a measuring fault. The difference is real.
Yes the deviation is not too much but before calibration it was perfect !

I did also a measurement with a level calibrator output before upgrade and after upgrade + calibration.
Have a look at the Pictures...

[f1rmb]

Hi,

Hi David,

Averaging  or Normal: No difference
With bandwidth limit: No difference (some times 1mV)
I checked the measurement also with my Racal Dana TRUE  Level Meter.
The difference is stable. (I know both meters are accurate)
It is not a measuring fault. The difference is real.
Yes the deviation is not too much but before calibration it was perfect !

I did also a measurement with a level calibrator output before upgrade and after upgrade + calibration.
Have a look at the Pictures...

    Just a little question, what kind of probe are you using ? (I'm seeing 50 Ohms input impedance selected in the screenshot).


Cheers.
---
Daniel

[Neddie]

Any chance the calibration has to be performed at a specific temperature?
Neddie

[Willi100]

Right
It's all 50 Ohm.
The probes (at the Level-Meters) are special RF-Power sensor heads with 50 Ohm.
You need it that you can measure the right RF-Level.

Willi

Edit: From the Signal-Generator there is a (50 Ohm) koax-kable direct to the scope input with 50 Ohm.

[Willi100]

@ Neddie
I play with that idea at the moment.
But It's not a big deal....
The much better way would be to change the intern voltage reference.

Willi

[JAmerican]

@ Neddie
I play with that idea at the moment.
But It's not a big deal....
The much better way would be to change the intern voltage reference.

Willi

Any updates on this? Anyone else experience this issue?

[JDubU]

I've noticed that the self calibration of my DS2072 seems to be better when I do it in an isolated, electrically quiet environment. 
I was initially running the self calibration within a few feet of other equipment (e.g. computer and monitor) and the baselines would end up being randomly off by a tiny bit relative to the baseline markers.

[miguelvp]

Also allow for the device to warm up.

I know my 2nd Channel takes longer to settle than Chan 1 for some reason.